The therapeutic range for efavirenz plasma concentrations is unclear and some studies found no correlation with viral non-suppression. Efavirenz concentrations are variable, driven in part by polymorphisms in CYP2B6. We hypothesised that efavirenz mid-dosing concentrations, together with CYP2B6 metaboliser genotype, could predict viral non-suppression. Participants starting first-line efavirenz-based antiretroviral therapy were monitored for 48 weeks. HIV-RNA and efavirenz mid-dose interval concentrations were determined at Weeks 16 and 48. CYP2B6 metaboliser genotype status was determined by 516G→T and 983T→C polymorphisms. Cox proportional hazards modelling was used to predict viral non-suppression and to determine the most predictive efavirenz mid-dosing concentration threshold. In total, 180 participants were included. Median efavirenz concentrations were 2.3 mg/L (IQR 1.6–4.6 mg/L) and 2.2 mg/L (IQR 1.5–3.9 mg/L) at Weeks 16 and 48, respectively. Moreover, 49 (27.2%), 84 (46.7%) and 39 (21.7%) participants had extensive, intermediate or slow CYP2B6 metaboliser genotype, respectively. Log2 efavirenz concentrations [adjusted hazard ratio (aHR) = 0.77, 95% CI 0.67–0.89] and baseline CD4 cell count (aHR = 0.994, 95% CI 0.989–0.998), but not CYP2B6 genotype, were predictive of viral non-suppression. For every doubling of efavirenz concentration there was a 23% decrease in the hazard of non-suppression. A threshold of 0.7 mg/L was found to be the efavirenz mid-dosing concentration that was most predictive of non-suppression. Mid-dosing efavirenz concentrations are predictive of viral non-suppression, but the currently recommended lower therapeutic limit (1 mg/L) is higher than our finding. Knowledge of CYP2B6 metaboliser genotype is not required for prediction of virological outcomes.

At the time of writing, David Bangsberg was affiliated with Harvard Medical School and Massachusetts General Hospital.